Mark McCarron
banner
mmccarron.bsky.social
Mark McCarron
@mmccarron.bsky.social
Computer Scientist
Psychological warfare's mathematical foundations. This isn't your grandpa's leaflet campaign. Learn Voodoo.

yetagainanotherhack.wordpress.com/2025/05/05/v...
Voodoo: The Science of Psychological Operations
PsyOps in the context of warfare is typically defined as any action which has, as its intent, the goal of altering or reaffirming the psychology of an enemy to the benefit of an attacker.  Typ…
yetagainanotherhack.wordpress.com
October 3, 2025 at 8:56 PM
Pain is something that is. A distinct aspect of reality. LLMs merely look up a higher dimensional model. There is no relationship between the two.
Also the big problem with ascribing pain to an LLM is that all neurons are effectively equivalent and that expressions of pain are little more than a shuffling of tokens - if any of these produce an actual sensation of pain it's likely because all activations does, because the nodes are pathological
October 3, 2025 at 8:52 PM
Reposted by Mark McCarron
Tree rings show Earth was hit by an extreme solar storm 14,300 years ago, likely caused by multiple CMEs striking in succession. It flooded the atmosphere with radiation, far beyond modern levels, and shows how intense solar activity can become.
🔭🧪⚒️
#SolarStorms
#SpaceWeather
#Paleoclimate
The Most Violent Solar Storm Ever Detected Hit Earth in 12350 BCE
A huge cosmic event some 14,300 years ago was so powerful, it left a discernible mark on our planet.
www.sciencealert.com
May 17, 2025 at 7:43 PM
This article takes a brief look at the #AI and #Robotics safety issues and demonstrates an ChatGPT jailbreak. This is a good piece if you want to understand the systemic issue at a deeper level.

yetagainanotherhack.wordpress.com/2025/02/11/a...
AI, Robots and the Dangers
AI safety has been a major theme over the last decade. In recent years, there has been much rolling back of these alignment studies. Every so often we hear of engineers leaving one company or anoth…
yetagainanotherhack.wordpress.com
March 1, 2025 at 11:01 PM
Reposted by Mark McCarron
I think modern computer software was a mistake and we should go back to our roots of developing for limited hardware.

The same amount of electricity that could power MS-DOS for a year is now used to ask Copilot how to save a Word document to your local Desktop instead of OneDrive.
January 28, 2025 at 8:34 PM
The problem with agents will be the accumulation of errors limiting the complexity of tasks which can be performed.
The true magic of #AI agents emerges when they, powered by
#LLMs, transcend static interactions and evolve into dynamic, decision-driven systems that revolutionize our connection with technology.
December 16, 2024 at 10:14 PM
Hallucinations in GenAI are caused by a malformed higher dimensional object.
December 16, 2024 at 9:49 PM
People who think consciousness will emerge in GenAI don't understand how Transformers work. There is no intelligence, instead there is a mathematical object which makes language and real-world relationships computable.
December 16, 2024 at 9:48 PM
What we need is a more formal approach to discovering the optimal higher dimensional object for a given language, rather than trying to construct it from data. This will probably be the next phase of AGI research.
December 16, 2024 at 9:48 PM
The real problem with GenAI and why hallucinations occur is because the higher dimensional object which captures the relationships is imperfect.
December 16, 2024 at 9:48 PM
I think it is time we move away from concepts such as learning when it comes to GenAI. GenAI represents information in higher dimensions which makes the problem, to a degree, linearly separable.
December 16, 2024 at 9:47 PM
GR's deep symmetries and the mathematical structures of relativistic theories serve as a form of intrinsic hypervisor. #physics
December 9, 2024 at 7:14 PM
An article to make you pause and think about life for a while. We don't do this often enough, instead we mainly put the blinkers on continue on with our day. Take a existential walk off the deep end with me.

yetagainanotherhack.wordpress.com/2024/12/01/t...
The Matrix: Going Deep
Ever thought there is something wrong with the world? Something which is like a thorn, but you can’t quite put your finger on it? You are not alone. I have covered some of this in previous ar…
yetagainanotherhack.wordpress.com
December 1, 2024 at 1:06 AM
AI isn't just learning from humans; it’s reflecting our biases, aspirations, and fears. In many ways, AI is a mirror. What is #AI like without the guardrails? What would that say about us?
November 28, 2024 at 8:04 PM
What happens when machines create things we don’t understand? #AI can write poetry or code, but does it mean anything to itself?
November 28, 2024 at 8:03 PM
Currently available LLMs are more intelligent than the vast majority of people, especially when we consider the breadth of knowledge they have. What you are really get at is a system which has its own agency and that's not possible with computational systems.
If we succeed in developing artificial superintelligence, what are the odds that it could turn into a god or a satan?
November 23, 2024 at 5:55 PM
But, they were partially correct. The printing press played a key role in the largest scale conflicts and mass deaths the world has ever known. #AI has the potential to do the same. Fact checking may not resolve the issue. Take Eugenics for example, scientifically correct, but morally wrong.
When the printing press emerged in the 1400s, critics feared it would weaken memory, spread dangerous ideas and make books less reliable than hand-copies. Today's resistance to #AI and #LLMs echoes those same fears - as if knowledge becoming more accessible somehow makes it less valuable.
November 23, 2024 at 4:21 PM
Still one of my favourites...

www.youtube.com/watch?v=1O_I...
Kenya Grace - Strangers (Kanine Remix)
YouTube video by Kanine
www.youtube.com
November 23, 2024 at 3:40 PM
Reposted by Mark McCarron
🚀 Who says you need a supercomputer for #AI? Meet #BabyGPT - a language model trained on a single laptop by our senior researcher Dr Amjad Majid.

Read more: bit.ly/4fvHQR7
GenAI Innovation: How Far Can a Single Laptop Take Us
Martel proves that powerful AI is possible on laptops. Training language models on a single laptop is possible, demonstrates Martel scientist.
bit.ly
November 22, 2024 at 11:21 AM
Reposted by Mark McCarron
Scaling alone won't be enough to achieve #AGI.
November 17, 2024 at 2:41 AM
Discover what it will take to make the leap to #AGI / #ASI from
#LLMs. This article will take you on a deep dive into the world of strong #AI

yetagainanotherhack.wordpress.com/2024/11/15/f...
From LLM To AGI: A Deep Dive
With LLMs the star of the AI world at the moment, I thought we should take some time out and discuss the differences between them and what it takes to deliver true AGI. LLMs take as an input some s…
yetagainanotherhack.wordpress.com
November 16, 2024 at 11:04 PM
A deep remix of a classic...
www.youtube.com/watch?v=okVE...
Sub Focus - X-Ray (Metrik Remix)
YouTube video by UKF Drum & Bass
www.youtube.com
November 16, 2024 at 10:18 PM
One of my favourites...
www.youtube.com/watch?v=lWc-...
Kanine - Light It Up
YouTube video by UKF Drum & Bass
www.youtube.com
November 16, 2024 at 10:09 PM